• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3772-3790
• /
• 2016

It is well known that the constant modulus algorithm (CMA) presents a large steady-state mean-square error (MSE) for high-order quadrature amplitude modulation (QAM) signals. In this paper, we propose a low-complexity hybrid adaptive blind equalization algorithm, which augments the CMA error function with a novel constellation matched error (CME) term. The most attractive advantage of the proposed algorithm is that it is computationally simpler than concurrent CMA and soft decision-directed (SDD) scheme (CMA+SDD), and modified CMA (MCMA), while the approximation of steady-state MSE of the proposed algorithm is same with CMA+SDD, and lower than MCMA. Extensive simulations demonstrate the performance of the proposed algorithm.

• Journal of the Korean Society for Precision Engineering
• /
• v.21 no.2
• /
• pp.130-137
• /
• 2004

A dynamic model of railway reducer is developed by the lumped parameter method. The model accounts for shafts, bearings flexibilities, gyroscopic effects and the force couplings among the transverse and torsion motions due to gearing. Vibration/noise analysis as well as strength of gear teeth, and bearing life are considered. Excitation forces of railway reduction are considered as the mass unbalance of the rotors, misalignment and a function of gear transmission error which comes from the modified tooth surface. A campbell diagram, in which the excitation sources caused by the mass unbalance of the rotors, misalignment and the transmitted errors of the gearing are considered, shows that, at the operating speed, there are not the critical speed. The program which can be used to analyze and predict vibration/noise characteristics by mass unbalance, misalignment and gear transmission error of railway reduction is developed with this system model.

• Journal of the Ergonomics Society of Korea
• /
• v.16 no.2
• /
• pp.1-14
• /
• 1997

In conventional probability-based human reliability analysis, the basic human error rates are modified by experts to consider the influences of many factors that affect human reliability. However, these influences are not easily represented quantitatively, because the relation between human reliability and each of these factors in not clear. In this paper, the relation is expressed quantitatively. Furthermore, human reliability is represented by error possibilities proposed by Onisawa, which is a fuzzy set on the interval [0,1]. Fuzzy reasoning is used in this method in order to obtain error possibilities. And, it is supposed that many basic events affected by the above factors are connected to the top event through Fault Tree structure, and an estimate of the top event expressed by a member- ship function is obtained by using the fuzzy measure and fuzzy integral. Finally, a numerical example of human reliability analysis obtained by this method is given.

• Journal of Ocean Engineering and Technology
• /
• v.10 no.1
• /
• pp.65-74
• /
• 1996

Up to now, the existing crack growth modelling has used a mathematical approximation but an assumed function have a great influence on this method. Especially, crack growth behavior that shows very strong nonlinearity needed complicated function which has difficulty in setting parameter of it. The main characteristics of neural network modelling to engineering field are simple calculations and absence of assumed function. In this paper, after discussing learning and generalization of neural networks, we performed crack growth modelling on the basis of above learning algorithms. J'-da/dt relation predicted by neural networks shows that test condition with unlearned data is simulated well within estimated mean error(5%).

• Atmosphere
• /
• v.22 no.2
• /
• pp.233-243
• /
• 2012

The Barnes scheme is used in Digital Forecast System (DFS) of the Korea Meteorological Administration (KMA) for real-time analysis. This scheme is an objective analysis scheme with a distance-dependent weighted average. It has been widely used for mesoscale analyses in limited geographic areas. The isotropic Gaussian weight function with a constant effective radius might not be suitable for certain conditions. In particular, the analysis error can be increased for stations located near mountains. The terrain of South Korea is covered with mountains and wide plains that are between successive mountain ranges. Thus, it is needed to consider the terrain effect with the information of elevations for each station. In order to improve the accuracy of the temperature objective analysis, we modified the weight function which is dependent on a distance and elevation in the Barnes scheme. We compared the results from the Barnes scheme used in the DFS (referred to CTL) with the new scheme (referred to EXP) during a year of 2009 in this study. The analysis error of the temperature field was verified by the root-mean-square-error (RMSE), mean error (ME), and Priestley skill score (PSS) at the DFS observation stations which is not used in objective analysis. The verification result shows that the RMSE and ME values are 1.68 and -0.41 in CTL and 1.42 and -0.16 in EXP, respectively. In aspect of spatial verification, we found that the RSME and ME values of EXP decreased in the vicinity of Jirisan (Mt. Jiri) and Taebaek Mountains. This indicates that the new scheme performed better in temperature verification during the year 2009 than the previous scheme.

• Proceedings of the Korean Institute of IIIuminating and Electrical Installation Engineers Conference
• /
• 2004.11a
• /
• pp.179-181
• /
• 2004

본 논문에서는 변형 삼각간섭계를 이용한 간섭패턴 생성과 관련한 몇 가지 실험결과를 제시한다. Two-pupil 합성 방법을 이용하여 변형 삼각간섭계의 광전달함수를 합성하고 광전달함수에 내포된 오차에 대한 원인을 분석해 보았다.

• Journal of Communications and Networks
• /
• v.6 no.3
• /
• pp.209-215
• /
• 2004

In this paper, a butterfly Log-MAP decoding algorithm for turbo code is proposed. Different from the conventional turbo decoder, we derived a generalized formula to calculate the log-likelihood ratio (LLR) and drew a modified butterfly states diagram in 8-states systematic turbo coded system. By comparing the complexity of conventional implementations, the proposed algorithm can efficiently reduce both the computations and work units without bit error ratio (BER) performance degradation.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.2
• /
• pp.189-194
• /
• 2020

This paper analyzes the gradient descent method, which is the one most used for learning neural networks. Learning means updating a parameter so the loss function is at its minimum. The loss function quantifies the difference between actual and predicted values. The gradient descent method uses the slope of the loss function to update the parameter to minimize error, and is currently used in libraries that provide the best deep learning algorithms. However, these algorithms are provided in the form of a black box, making it difficult to identify the advantages and disadvantages of various gradient descent methods. This paper analyzes the characteristics of the stochastic gradient descent method, the momentum method, the AdaGrad method, and the Adadelta method, which are currently used gradient descent methods. The experimental data used a modified National Institute of Standards and Technology (MNIST) data set that is widely used to verify neural networks. The hidden layer consists of two layers: the first with 500 neurons, and the second with 300. The activation function of the output layer is the softmax function, and the rectified linear unit function is used for the remaining input and hidden layers. The loss function uses cross-entropy error.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.7A
• /
• pp.520-528
• /
• 2003

This paper proposes a method of implementing synthetic stereo and an acoustic echo cancellation algorithm for multiple participant conference system. Synthetic stereo is generated by HRTF and two loudspeakers. A robust adaptive algorithm for synthetic stereo echo cancellation is proposed to reduce the weight misalignment due to near-end speech signals and ambient noises. The proposed adaptive algorithm is modified version of SMAP algorithm and the coefficients of adaptive filter is updated with cross correlation of input and estimation error signal normalized with sum of the autocorrelation of input signal and the power of the estimation error signal multiplied with projection order. This is more robust to projection order and ambient noise than conventional SMAP. Computer simulation show that the proposed algorithm effectively attenuates synthetic stereo acoustic echo.

• Journal of Institute of Control, Robotics and Systems
• /
• v.7 no.8
• /
• pp.703-710
• /
• 2001

The nonlinear observers are proposed for a nonlinear system. To improve the characteristics such as stability, convergence, and $H^{\infty}$ filter performance criterion, we utilize an $H^{\infty}$ filter Riccati equation or a modified $H^{\infty}$ filter Riccati equation with a freedom parameter. Using the Lyapunov function method, the characteristics of the observers are analyzed. Then the in-flight alignment for a strapdown inertial navigation system(SDINS) is designed using the proposed observer. And the additive quaternion error model is especially used to reduce the uncertainty of the SDINS error model. Simulation results show that the observer with the modified $H^{\infty}$ filter Riccati equation effectively improves the performance of the in-flight alignment.